collecting sales data for goods or services from a retail outlet;

transmitting the sales data to a third party through an open-architecture computer network and storing the sales data on a hard disk;

retrieving the sales data from the hard disk;

providing a likelihood function as a function of the sales data, base demand parameters for the goods or services, and time-varying demand parameter for the goods or services, wherein the likelihood function is a Gaussian-based likelihood function which includes a control parameter as a function of multiple time periods of the time-varying demand parameter for the goods or services;

solving, by a computer processor, for the base demand parameters for the goods or services and time-varying demand parameter for the goods or services using an inverse Hessian with tri-diagonal band matrix (TDBM), the inverse Hessian with TDBM having horizontal and vertical bands for the base demand parameters for the goods or services and a diagonal band for the time-varying demand parameter for the goods or services;

providing, by the computer processor, a non-stationary time series model from an expression using the solution of the base demand parameters for the goods or services and time-varying demand parameter for the goods or services, the non-stationary time series model including a reference demand profile for the goods or services incorporating known events from the retail outlet within the non-stationary time series model, the known events including holiday spikes;

transmitting the non-stationary time series model to the retail outlet through the open-architecture computer network, the non-stationary time series model being displayed via a website; and

controlling, by the computer processor, operation of the retail outlet by setting a price for the goods or services based on the non-stationary time series model.